Climate Science Glossary

Term Lookup

Enter a term in the search box to find its definition.


Use the controls in the far right panel to increase or decrease the number of terms automatically displayed (or to completely turn that feature off).

Term Lookup


All IPCC definitions taken from Climate Change 2007: The Physical Science Basis. Working Group I Contribution to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Annex I, Glossary, pp. 941-954. Cambridge University Press.

Home Arguments Software Resources Comments The Consensus Project Translations About Donate

Twitter Facebook YouTube Pinterest

RSS Posts RSS Comments Email Subscribe

Climate's changed before
It's the sun
It's not bad
There is no consensus
It's cooling
Models are unreliable
Temp record is unreliable
Animals and plants can adapt
It hasn't warmed since 1998
Antarctica is gaining ice
View All Arguments...

Keep me logged in
New? Register here
Forgot your password?

Latest Posts


Climate Hustle

Phil Jones - Warming Since 1995 is now Statistically Significant

Posted on 14 June 2011 by dana1981

As all Skeptical Science readers are undoubtedly aware, in February of 2010, Phil Jones was asked some loaded questions in an interview with the BBC.  Several of the questions were gathered from "climate sceptics", and Jones' answer to the second one has been widely re-published and distorted:

"Do you agree that from 1995 to the present there has been no statistically-significant global warming?"

"Yes, but only just. I also calculated the trend for the period 1995 to 2009. This trend (0.12C per decade) is positive, but not significant at the 95% significance level. The positive trend is quite close to the significance level. Achieving statistical significance in scientific terms is much more likely for longer periods, and much less likely for shorter periods."

Why choose 1995 as the starting point in this question?  Well, that is the closest year for which the answer to this loaded question is "yes".  From 1994 to 2009, the warming trend in the HadCRUT dataset was statistically significant at the 95% confidence level (CL).  It's also worth noting that there's nothing magical about the 95% CL - it's simply the most commonly-used interval in scientific research, but it's also true that the HadCRUT 1995-2009 trend was statistically significant at a 93% confidence level. 

In other words, using Jones' data, we could say with 93% confidence that the planet had warmed since 1995.  Nevertheless, this did not stop numerous mainstream media outlets like Fox News claiming that Phil Jones had said global warming since 1995 was "insignificant" - a grossly incorrect misrepresentation of his actual statements.  The Daily Mail warped the truth even further, claiming Jones had said there was no global warming since 1995.  These media outlets turned 93% confidence of warming into "no warming". 

Furthermore, the HadCRUT dataset excludes portions of the Arctic where there are no temperature stations.  The Arctic also happens to be the fastest-warming part of the planet.  NASA's GISTemp, whose data analysis extrapolates for the Arctic temperatures using the nearest temperature stations, did find a statistically significant warming trend at the 95% CL from 1995 to 2009.  So not only are the "skeptics" cherrypicking the start date, they're also cherrypicking a dataset which doesn't cover the whole planet.

Deep Climate has detailed the history of the 1995 cherrypicked starting point.  It appears to have originated with an email from Richard Lindzen to Anthony Watts, which was subsequently published in a post on WattsUpWithThat (WUWT):

Look at the attached.  There has been no warming since 1997 and no statistically significant warming since 1995.  Why bother with the arguments about an El Nino anomaly in 1998?  (Incidentally, the red fuzz represents the error ‘bars’.)

Best wishes,


Luboš Motl made a similar argument in December 2009 using UAH satellite data, which was also published on WUWT.  Two months later, the question was posed to Phil Jones in the BBC interview, which suggests strongly that it originated from Motl, Lindzen, and/or Watts.  Regardless of the source, what really matters is that the question was based on a cherrypicked starting date, and on a somewhat arbitrary statistical confidence level, and that the media subsequently distorted Jones' response.

In January 2009, Tamino at Open Mind analyzed the data after removing the influence of exogenous factors like El Niño, volcanic eruptions, and solar variation from the temperature data.  Tamino concluded that "until 2001 the warming is statistically significant" (Figure 1).

tamino analysis

Figure 1: HadCRUT3v estimated warming rates from the plotted date to Present with 2-sigma error bars, using exogenous factor-compensated temperature data (Open Mind)

Another year has passed since the original BBC interview, and in a new BBC article, Jones notes that the HadCRUT warming trend since 1995 is now statistically significant.

"Basically what's changed is one more year [of data]. That period 1995-2009 was just 15 years - and because of the uncertainty in estimating trends over short periods, an extra year has made that trend significant at the 95% level which is the traditional threshold that statisticians have used for many years.

"It just shows the difficulty of achieving significance with a short time series, and that's why longer series - 20 or 30 years - would be a much better way of estimating trends and getting significance on a consistent basis."

As Jones notes, and as scientists like Lindzen and Motl should very well know, trying to assess trends in the noisy global temperature data over periods as short as 10-15 years is pointless.  There's just too much short-term noise, which if you're going to look at such short-term data, you at least need to attempt to filter out first, as Tamino did.

So to sum up, a cherrypicked starting date chosen by a couple of "skeptics" (Lindzen and Motl) and published by a "skeptic" blog (WUWT) was picked up and passed along in the form of a loaded question to Phil Jones in the BBC interview.  Phil Jones' answer was subsequently (and predictably) grossly distorted by various media outlets, who turned 93% confidence of global warming into "no global warming". 

In reality, the HadCRUT warming trend since 1995 was statistically significant above the 90% CL, the GISTemp warming trend (which does not exclude the Arctic) was significant at the 95% CL, and by removing short-term effects, even HadCRUT has been significant at the 95% CL since 2000.  One year later, we can now say that the HadCRUT warming trend since 1995 is statistically significant at the 95% CL, even including the exogenous factors.

Unfortunately, the main consequence of this sequence of events was that much of the public was misinformed by media articles claiming that global warming since 1995 was "insignificant" or non-existent, which are both factually incorrect statements.  Misleading the public may well have been the goal of those individuals who originally cherrypicked the 1995 starting date and the HadCRUT dataset, and if so, they succeeded.  And not surprisingly, Anthony Watts continues to mislead his readers, claiming Phil Jones' comments are "an about face...From the “make up your mind” department", when in reality Jones' comments have been consistent and accurate throughout.

This reactions to this story have revealed a number of media outlets whose aim is not to accurately inform their readers with regards to the climate, but rather to misinform them.

0 0

Bookmark and Share Printable Version  |  Link to this page


Prev  1  2  3  4  Next

Comments 101 to 150 out of 181:

  1. MoreCarbonOK ROTFLMAO, as it happens my field of research is a branch of statistics, and I have a long list of peer-reviewed journal articles on statistical methodology, and so I find your hubris rather amusing!

    Do note however that I have pointed out two specific flaws in your analysis, and for all of your bluster, you haven't even attempted to address either of them. That speaks volumes for the strength of your analysis.
    0 0
  2. Dikran: perhaps a better answer would be along the lines of, "Over that interval, the warming trend is only 93% significant, which doesn't meet our usual standard of 95% significance."
    Correct, answering the question (if slightly indirectly), and with a much better chance of letting the audience know it's only *just* failing the significance test, but the trend is definitely still upwards.
    Of course, you'd say that would be reported as "[it] doesn't meet our... standard", and you'd probably be right, too. But it'd be a lot harder to spin it into "Phil Jones says it hasn't warmed since 1995", given that there's a clear statement that it *is* warming. Pointing back to the original quote would highlight the amount of word-twisting that had to go on to get the desired headline.
    0 0
  3. Bern, that answer is better than most, but again what happens if the interviewer thinks it is too indirect and presses for a definitive answer. Then you end up having to say "no it is not significant", but you have made it look as if you had made a concession and had been evasive.

    Note it is also pretty much what Jones actually said, other than having started by giving a direct answer.

    Nothing Jones actually said implies that it hadn't warmed since 1995, but the Daily Mail reported it that way anyhow. Note Jones actually said a bit later in the interview "I'm 100% confident that the climate has warmed", so the quote is there.

    N.B. If he said specifically that it had warmed since 1995 he would again be accused of scientific malpractice as you are not supposed to claim that something exists based on a failed test of statistical significance. It is the fault of hypothesis testing, which is rather broken in the way it is applied in science.
    0 0
  4. MoreCarbonOK, I can't find Easterling's 2005 paper online. Here's a paper that references that paper and uses some of that data (showing minimum temperatures rising faster than maximums):
    0 0
  5. @Eric
    The link reports that the file is damaged.

    I should report that it has not happened often on my pool table that minima were rising higher except for in the case of Honululu, which, as I said before, could be due to volcanic action. The minima in the arctic (Norway) was also slightly higher but only just. All in all, the maxima rose a lot faster than minima. And that means the warming is natural.

    Dikran says: I have a long list of peer-reviewed journal articles on statistical methodology, and so I find your hubris rather amusing

    (-Inflammatory and ideology snipped-).
    0 0
    Moderator Response: (DB) Please take care to formulate your comments to better comply with this site's comment policy.
  6. Henry@Bern

    I am sure my results are repeatable, but it is a lot of work. Don't think you can do it in 5 minutes. Forget about confidence intervals, you re. to it means you have not understood my method.

    You have to get access to the original reported data and plot the average reported monthly values against time.
    You can do that in Excel, which affords you to quickly get the trendlines.

    As an example, here is one of the first stations that I evaluated. It is an island that lies south of South Africa.
    this is how you must do it.
    0 0
  7. MoreCarbonOK, I checked your site and the source of your data, which is not an official source. To compare the data with official sources, I checked the Australian Bureau of Meteorology (BoM) statistics for Brisbane Airport.

    The clear conclusion is that your analysis of Brisbane temperatures is wrong. To what extent this is due to dodgy data, and to what extent due to statistical errors as identified by Dikran Marsupial I cannot say.

    The first thing I noticed is that the station identification numbers do not match. Nor is the station identification number used by your source used by BoM for any station. Further, there are slight differences in the specified location as to longitude and latitude. Checking nearby stations, I determined that your site was closer to Brisbane Airport than to other nearby BoM sites (Cape Moreton, Brisbane City, Archerfield and Amberly). So if the numbers you have come from an official BoM site, they do come from Brisbane Aerport.

    I then proceeded to check the daily values for May, 2011. I noticed that your figures differ from the BoM figures with regard to Minimum Temperature and Relative Humidity for May 1st; on Relative Humidity for May 2nd; Minimum Temperature and Relative Humidity on May 3rd; and on Minimum Temperature for May 4th. I did not bother checking any further. For just 12 values, your sight differed from BoM figures for 6 of those values. It is a reasonable presumption, therefore, that your analysis is based on faulty data and cannot be assumed to accurately reflect reality.

    For the record, BoM calculates trends and means for Brisbane Airport. For the period 1910 to 2010, the the mean temperature shows a positive trend of 0.04 degrees C per decade; the Minimum Temperature shows a trend of 0.11 degrees C per decade, and the Maximum temperature shows a trend of - 0.02 degrees C per decade. Almost all of the cooling in the Maximum trend was between 1910 and 1960, with the trend in the last half few decades of the twentieth century being positive. The greatest trend in the mean temperatures has also been in that period, but the greatest trend in minimum temperatures was in the period 1960 to 1990. Obviously the diurnal temperature range has narrowed over the preceding century (-0.13 C per decade) although recently it has widened noticeably.
    0 0
  8. Dikran and others, I fear you are wasting your time with MoreCarbonOK - you only have to look at his website, to see classics like this :

    A short summary of the basic results of my study:
    CO2 is insignificant as a greenhouse gas.
    CO2 is not a poison or a pollutant.
    CO2 is one of the two main building blocks of ALL plant life on Earth.
    CO2 concentration has been up to ten times higher in the past
    CO2 is good.

    To start off with, I found Svante Arrhenius' formula completely wrong...

    It appears that we have not even yet reached the level of warmth compared to previous warm periods in known history, such as for example, the Medeviel Warm Period (MWP) - 1000 years ago, when Greenland was really green. See here:
    (Link to WUWT removed)

    Now the opposite, global cooling, that really is something to worry about....

    There are some people there who make admirable attempts at showing him where he is going wrong, but there are none so deep in denial as those who believe that they have the secret answer and who believe they are more intelligent than the rest of us mere mortals !
    0 0
  9. I think I would have said something along the lines of:

    "The 'statistical significance' test is 95% confidence, which is to say one chance in 20 that your data shows the trend incorrectly, that it's instead your null hypothesis plus noise.

    For that 15 year period the confidence is ~93%, or about one chance in 14 that the data doesn't correctly reflect the trend, so while it's suggestive, that short time frame doesn't meet the generally used significance criteria."
    0 0
  10. Sorry, misplaced my second paragraph in 107 while editing. It should be placed at the end of the above comment (where it makes more sense).
    0 0
  11. Dikran Marsupial @77, certainly the deniers could have misrepresented Jones as you indicate if he had replied as I suggest. The issue is not whether what he said can be misrepresented, as everything can. The issue is how clear is it that it is a misrepresentation for those seeing the initial quote. In the instance where deniers had tried your suggested response, they leave themselves open to the obvious counter response that of course it is inappropriate to base claims on data that is not statistically significant; that that is why Jones bases his claims on thirty year trends which most certainly are warming, and are statistically significant; but that he was asked about a particular period which was only being discussed because deniers cherry picked a short interval to misrepresent the actual trend. The issue has then become the thirty year trend, and denier cherry picking rather than obscure arguments about which definition of "significance" is being used.

    Having said that, while I do believe that Jones could have answered better, I am in no way inclined to condemn him for that. As can be seen by this discussion, it is by no means easy to formulate an answer that is difficult to misrepresent, and far more so when doing so on the fly. And as I previously noted, and contrary to John Russell, even pausing to think can be misrepresented by hostile editing, so answering on the fly is more or less necessary.

    Where I do part ways with you, and agree with John Russell is that because it is always possible to answer better, I think media training would be useful for climate scientists who are often called upon to front the media.
    0 0
  12. KR sadly that interpretation isn't correct. Statistical significance at the 95% level means that there is a 1 in 20 chance of seeing a trend as large or larger assuming the null hypotheis is true. That does not mean that there is a 1 in 20 chance that the null hypothesis actually is true. This is mixing frameworks, but the p-value is p(x>X|H0), where x is the trend and X is the observed trend and H0 is the event that the null hypothesis being correct, however what you actually want is p(H0|x>X), which you can get by Bayes formula

    p(H0|x>X) = p(x>X|H0)p(H0)/(p(x>X|H0)p(H0) + p(x>X|H1)p(H1))

    however the hypothesis test does not involve p(H0), p(H1) (the prior probability that the null and alternative hypotheses are correct) and p(x>X|H1), which measures how likely such an extreme statistic is assuming that there is a secular trend.

    Essentially a frequentist hypothesis test does not allow you to assign a numeric probability to either hypothesis being correct as it can only assign probabilities to things that have long run frequencies and the a particular hypothesis is either true or ot isn't. However, that is exactly the form of answer we actually want, which is why the p-value is so often misinterpreted that way. That makes it very difficult to give an accessible answer to the question that is also correct as statistical significance simply doesn't mean what people think it means.

    JMurphy Absolutely, I gave him the benefit of the doubt, but he showed he was a denialist by being unwilling to engage with the substantive technical criticism of his methodology. The hubris was priceless, though ;o)
    0 0
  13. Tom Curtis wrote "Where I do part ways with you, and agree with John Russell is that because it is always possible to answer better, I think media training would be useful for climate scientists who are often called upon to front the media. "

    Actually, I don't disagree with that at all, in fact I said media training is vital for any scientist that needs to communicate with the media. It is the particular question that is the problem, I doubt there is an accessible answer that can't be substantially misconstrued. I think Jones' answer was better than any of the alternatives presented so far. The was a logical disconnect between the "the trend is not statistically significant" and "there has been no warming". If your opponent is not required to be logically correct, there is little you can do, especially with a topic as counter-intutive as hypothesis testing. There are plenty of working scientists who don't understand it properly.
    0 0
  14. Dikran - So, incorrect terminology and implications (thanks, I don't have multiple publications in statistics):

    Would this be more correct?

    "Statistical significance at the 95% level means that there is a 1 in 20 chance of seeing a trend as large or larger assuming the null hypotheis is true, while statistical significance at the 93% level means that there is a 1 in 14 chance of seeing a trend as large or larger assuming the null hypotheis is true.

    So again, while 93% significance over that period is suggestive, the short time frame data does not meet the statistical significance threshold, which is why 20-30 years of data is preferred to clearly identify trends."

    The major point aimed for here is to emphasize that while the data is pointing in a particular direction, it's not doing so strongly enough to meet the standard test - hopefully not leaving room for the interviewer (or onlookers) to say "No warming since 1995"...
    0 0
  15. JMurphy @108,

    The Moniker was a clear give away. IMHO, his posts belong on the "denial" thread more than they do here. And the irony, a post about cherry picking and the guilty party who thinks that doubling or trebling CO2 is OK goes right ahead an cherry picks to continue to reinforce their opinion on the matter.

    Meanwhile for the period which we almost have (almost complete global coverage) sufficient data to determine statistically significant temperature trends we get this:


    The planet is in a net energy imbalance and is with it folks.
    0 0
  16. Dikran writes: "Media training is vital for any scientist that need to have contact with the press, but nothing can provide a 100% bulletproof defense."

    Well then, I think we agree; because I was never saying there is a bullet proof way to respond. Rather I was saying that, in his innocence, Phil Jones made it much too easy and there would have been ways of expressing an honest answer that would have been more difficult to twist -- or even if twisted, easier to demonstrate as obfuscation. I'm sorry I can't come up with one, but that's because I'm not an expert on the science. Cheers!
    0 0
  17. MoreCarbonOK

    Pardon my earlier assumption that your wild variation of month based trends was due to your treatment. I was able to replicate a very similar trend of -0.108 C/yr for the Brisbane AP May Minima series. Although it is also apparent that the detrended standard deviation is 1.18 C (or 33% of the total trend change over the period).

    However the fact remains that, just as no-one would base average global temperatures on just 10 sites to show a statistically significant warming trend, no-one would seriously base a SH cooling trend on just 5 sites. You need far more site dirversity before you can make the point that you do.
    0 0
  18. KR yes, that interpretation is fine, but it still admits that it is not significant (so they would probably make the same claim anyway - logical consistency had nothing to do with it!).

    The real problem is then when a skeptic has a non-significant statistic, are they allowed to say that the statistic is suggestive of support for their theory? The reason we are not supposed to claim anything based on a statistically insignificant statistics is to guard against confirmation bias. Essentially we should always apply the intepretation that provides the weakest support for our own position. Jones did just that, the Daily Mail did the exact opposite and used the (incorrect) interpretation that provided the most support for their position. However in reality hypothesis tests generally provide little support for any position (and so a pessimistic intepretation is a good idea).

    I don't mean to be argumentative, it is just that Jones did a very good job of dealing with this very difficult question. I don't think I could have done any better without the use of diagrams and a page or two of text.
    0 0
  19. Dikran - "I don't mean to be argumentative, it is just that Jones did a very good job of dealing with this very difficult question. I don't think I could have done any better without the use of diagrams and a page or two of text."

    I agree. I find it extremely difficult to talk without a whiteboard sometimes.
    0 0
  20. I just think you have to keep the audience in mind. In a BBC interview, the audience is the general public. With that audience, you don't want to give a technical answer which is heavy on statistics and requires diagrams and several pages of text. Jones' answer was good if his audience was full of scientists and statisticians.

    I understand why Jones answered the way he did, and without giving it a lot of consideration before answering, I probably would have said "yes, but..." just like he did. But that's a mistake. I think you have to answer that question by explaining what's wrong with the question. You can't play into their trap by feeding them the answer they want.

    I think Jones should have used it as a teaching moment. "Well, here's why we look at temperature changes over periods of 20-30 years, and why 15 years is too short..." Then you can explain that it's not surprising why the 15 year trend is statistically insignificant, because you've already established that it's simply too short of a timeframe.

    As has been suggested, knowing how to properly answer this question probably requires some media training - learning how to communicate with the general public, who don't know what "statistically significant" means.
    0 0
  21. moreCarbonOK.... Reading through your comments here and looking through your site I believe I see most of your sources of information seem to be restricted to JoNova, WUWT, Monte Heib and others.

    With all due respect, you seem to be actively searching for sources that confirm what you want to believe. You've pieced together, for yourself, the same narrative being propagated by a very active anti-science crowd of people. The material you're putting together may feel right to you, it may be what you want to believe, it may sound very compelling to you. It's not science. It's not accurate.

    If you were truly skeptical you would take the time to try to learn what the actual scientific research says. The reason I've come to trust this particular blog (SkS) over many others is that it provides citations to actual research. Research that I go read for myself. In that I don't have to trust John Cook or any of the other authors here. I can be skeptical of even their claims. But they give me the opportunity to go to the source of the information and see for myself if their claims are accurate.

    I would urge you to try to do the same with JoNova, WUWT and others. I think you'll find that they mostly stick to referencing each other, not the published research. I suggest being skeptical of your own position. I try to do the same with my position every day.
    0 0
  22. Dana... You're totally right. I imagine Jones was torn between being accurate and feeling trapped by the question. I think he should have lead with a comment suggesting it's a misleading question. From the moment Jones said, "Yes. But only just." People only heard the word "Yes." Everything after that was just trying to climb out of the pit he'd fallen into. Better to point at the pit first, side step it, explain why it's a pit and then answer the question as accurately as possible.

    Hindsight 20/20.

    I can't find the link at the moment but the recent John Abraham radio interview was a very good example of side stepping the traps and accurately explaining the science.
    0 0
  23. Exactly Rob - the question was fed to Jones in a way to get him to answer "yes". As soon as he did, the "skeptics" got all they needed, as we saw from Fox News, Daily Mail, etc.

    I sympathize with Jones, because when I'm asked a yes or no question, I tend to answer yes or no and then clarify with the "but" if necessary. So like I said, my initial reaction would have been to respond the same way as Jones did. But when the question is being asked to trap you into a specific answer, you have to think outside the box and find a different way to answer it. And it's tricky, because scientists aren't normally asked loaded questions like that. Hence the need for media training.
    0 0
  24. 121 - Rob Honeycutt

    Thanks. I really was at a loss to see what MoreCarbonOK was trying to say. You seem to have been able to read his posts better than me.
    0 0
  25. FWIW, I would have answered something like this:

    "That's an interesting question, and let me tell you why it's an interesting question..."

    First align with the questioner, then differentiate to reveal the fallacy of the question.
    0 0
  26. Right Daniel, that's exactly what politicians do. Politicians have been exposed over the years to lots of actually good journalism, faced with questions to which they did not have a good answer, or no answer at all. They learned these methods to cope with it and come out at the end saying many words that add up to no meaning.

    It is rather ironic that now these methods should be used by honest people having to answer dishonest questions, but perhaps scientists should start familiarizing themselves with it, less their words be bent in endless twists of deception.
    0 0
  27. Philippe Chantreau IMHO that is exactly why few people would trust a politician further than they could throw them, the last thing scientists should do is follow in their footsteps!
    0 0
  28. @ Dikran: I'm not advocating to speak as a politician, but to first establish control of the narrative, then expose the question for the fallacy it is and to then give the actual answer with appropriate context to make it a teaching moment. But the initial sentence used is key, as it sets the stage for establishing control of the narrative to be played out. That initial response by Jones is the area of opportunity to seize control of a bad situation and to establish a more appropriate response. A good defense is to establish a counteroffensive which also dictates the terms of the engagement. Answering the question without reframing it is to engage at a disadvantage: the battle may ultimately be won but the campaign itself is lost.
    0 0
  29. To back up Daniel here... This is all just the nature of language and communication. How we structure our language is part of everyday life for all of us. Politicians use techniques to dodge questions in order to not answer them. I think what Daniel, Dana and I are suggesting is it's important to dodge the "trap" when a question is structured in a way that's intended for such a purpose. But also, follow that with a full and complete answer to the question.

    Phil Jones could clearly have answered the question in full without falling for the trap. I think Dr. Jones' Achilles heel is that he so much of a scientist that he answered the question in a manner, as Dana suggested, that addressed a scientific audience. Any scientist listening to his words knew exactly what he meant. The general public did not. And that was the trap.
    0 0
  30. I think Phil Jones gave an answer which was correct, easily understandable and honest. The fact that those in denial used a version of his answer to try to claim things that Phil Jones did not mean or say ("Global Warming in Last 15 Years Insignificant, U.K.'s Top Climate Scientist Admits" from FOX NEWS, and "Climategate U-turn as scientist at centre of row admits: There has been no global warming since 1995" from the DAILY MAIL - both headlines are demonstrably false), simply shows how despicable and dishonest they are.
    No matter how Jones had answered, his words would still have been twisted and abused.

    Don't have a go at Jones for being naive, credulous or whatever - his answer would never have been perfect enough to have disallowed its use as propaganda by those who thrive on disinformation. Have a go at those who misuse and misinform.
    0 0
  31. Daniel, the question isn't a fallacy, it is actually a perfectly reasonable question, it is the implicit conclusion that global warming is not happening that is a fallacy, but that was unsaid and was made afterwards by the "skeptic" media.
    0 0
  32. JMurphy... You're right. They absolutely would have spun the whole thing the way they wanted regardless of what Jones said. But, by opening his answer with the word "yes" he basically loaded their weapons for them.
    0 0
  33. @ Dikran

    I think we're in close agreement here. No matter the answer Jones gave, it was going to be spun out of context. My point I was trying to make (poorly) was that in not addressing the framing of the question immediately, anything that followed would be playing into the hands of the denialarati. That was the fallacy (perhaps there's a better word?).

    Thus, following a normal scientific response to a question given, the outcome had a predictable end.

    (It was essentially one of those "So, how long has it been since you stopped beating your wife?" questions)
    0 0
  34. @#95 MoreCarbonOK "I only later picked up on a trend of no warming in the SH and more warming in the NH. You tell me why. Anyone?"

    The land trends are higher than the sea surface trends. SH has higher percentage of ocean, therefore is expected to have a lower trend. Land trends are higher than sea surface trends because the heat capacity of land is lower than water.
    0 0
  35. This whole topic is a tempest in a teapot.

    It is strange, though, that Phil Jones uses suspect statistics to make the claim that the warming is now statistically significant at 95% confidence level.

    The HadCRUT3v (variance adjusted version) is not recommended for this sort of analysis.

    The time series shows autocorrelation (i.e. residuals are not white), so the uncertainty bands should be expanded. But he does not do so.

    Lucia's Blackboard has a more statistically accurate analysis of whether Phil Jone's statement is accurate. Note that she is consistent in widening the uncertainty bands ..... she also routinely goes through the different possible versions of uncertainty bands in here monthly comparison of GISS and HADCRUT3 vs the IPCC AR4 projections. In those cases, the widened uncertainty bands is the only thing that keeps GISS trendline from being statistically different than the IPCC projection.

    93% vs 94% vs 95% confidence factor isn't all that relevant and more likely than not, Phil Jones statement will be statistically valid in the next year or two. The "since 1995" is a cherry-picked starting point in the first place. But it is silly of him, however, to make this technically incorrect claim.
    0 0
  36. Charlie... Boy-o-boy did I have a knockdown drag out with Lubos Motl not long after the Phil Jones interview over whether anything less than 95% was real or not. Motl was absolutely adamant than anything less than 95% statistical significance, in fact, did not exist.

    Regarding Lucia... The old saying is, "There's more than one way to skin a cat." Tamino skin this cat quite differently and comes up with statistical significance from 1995-2009.
    0 0
  37. @Rob Honeycutt --- "Tamino skin this cat quite differently and comes up with statistical significance from 1995-2009"

    He didn't find statistical significance in either HadCRUT3 or HadCRUT3v. To quote Tamino "[Response: Only by removing the influence of exogenous factors (el Nino, volcanic, solar) can you show statistical significance post-1995.]"

    The "Tamino" series starts of with the HadCRUT3v, which Phil Jones himself says has artifacts from the gridding method that makes HadCRUT3 the recommended data for time series analysis. Tamino then applies his own recipe of modification to remove such things as solar influence. Then he does an analysis.

    The analysis is interesting, but doesn't support Phil Jones's statement.
    0 0
    Moderator Response: (DB) Actually, Tamino's analysis showed that the warming since 2000 and earlier was statistically significant.
  38. Charlie A @135, I am very far from expert on statistics, so I would appreciate somebody who is expert either confirming, or correcting the following (Hint: Dikran). However, it appears to me that your statement about that:

    "The time series shows autocorrelation (i.e. residuals are not white), so the uncertainty bands should be expanded. But he does not do so."

    is incorrect. Specifically, the time series does show autocorrelation, but the error bars show measurement error, which would be white noise. Consequently, the effect of auto-correlation in the series would be to reduce the expected error rather than increase it in that nearby terms would provide some (very limited) information about the probable values of their neighbours.
    0 0
  39. Curious thing, when Jones answered the loaded question and noted that the warming was not stat. significant at the 95% level, 'skeptics' then did not question his stats, or the particular data set that he used, in fact they accepted/endorsed the result without question and hyped the story. But now that he has determined that the warming in the HadCRUT3 (variance adjusted data) does meet the 95% criterion, they suddenly do take issue. Note too that they ignore the fact that the warming does meet the 95% confidence level in the GISTEMP data, not to mention all the other signs that the planet is warming:

    0 0
  40. @138 Tom Curtis.

    Without getting too heavily into the math, the effect of autocorrelation in the time series is to reduce the effective degrees of freedom, thereby increasing the size of the uncertainty band compared to the standard deviation of the samples.

    This also happens when one applies low pass filtering to a time series.

    The is not due to measurement error, but due to the long term persistence evident in virtually every climatological time series.
    0 0
  41. #139 Albatross "Note too that they ignore the fact that the warming does meet the 95% confidence level in the GISTEMP data"

    Phil Jones's statement was about HadCRUT time series, not GISS.

    The question is not even whether the climate is warming.

    The question is whether Phil Jones's statement true.

    There are many, many statements that Phil Jones could have made that would be true. He chose to make a statement not supported by the facts ..... at least if one does a proper statistical analysis.
    0 0
  42. Charlie,

    Re GISTEMP, read my post again (and my other posts on this thread), and carefully.
    0 0
  43. Charlie,

    "The question is whether Phil Jones's statement true"

    Sigh, trying very hard to reframe the argument and shift the goal posts are you not? Again, read my post above carefully. And also read the main post, the question that is being asked is this:

    "Why choose 1995 as the starting point in this question? "

    Also, have you asked yourself what type of statistical analysis and data Lindzen used to arrive at his conclusion that there had been no stat sig warming? Don't see you questioning his stats analysis, never mind the content of his email to Watts. He was showing Watts how to cherry pick and distort. Are you OK with that?

    I find your insinuations that Jones behave nefariously in this whole fiasco manufactured by Lindzen to be beyond the pale.
    0 0
  44. Charlie A (various posts). Yes, fine, the data are autocorrellated and an OLS trend is an optimistic calculation. However the basic assumptions of a hypothesis test are invalidated if you deliberately tune the start point to get the result you want (effectively it means you are performing a multiple hypothesis test, so the false-positive and false-negative error rates will be affecte dby that). The effects of not taking into account the autocorrellation is a minor issue compared to that. That means that just accounting for the autocorrellation would not be technically correct either.

    Secondly, it isn't a big deal anyway, simply because the expected magnitude of an anthropogenically induced trend is so small that you would not expect to reliably detect it over so small a window, as windows of that short duration tend to be dominated by sources of short term variabilty due to things like ENSO. See the paper by Easterling and Wehner which shows that GCM simulations show occasional periods of little or no warming, even during long term forced global warming of a decade or two, for precisely this reason. In other words the effective power of the test is very low.

    Lastly, a test of statistical significance provides evidence for making some claim based on the value of a single statistic alone. That is, if that one statistic were the only evidence you had that the climate was warming, you could only claim the planet was warming if there were a statistically significant trend. But the trend from 1995-2010 is not the only evidence we have, so Jones is perfectly justified in having 100% belief that the world is warming.

    Basically anyone who is making a big song and dance about the trend not being significant is defficient in their statistics for not understanding the multiple-hypothesis testing issue involved in the cherry picking of the start point to suit one particular argument, and defficient in their understanding of climate physics for not understanding why you wouldn't expect the trend to be statistically signficant with high probability even if there is a secular trend of the expected magnitude.
    0 0
  45. Charlie A - if you want to see a real error, then try Dick Lindzens comment "Look at the attached. There has been no warming since 1997 and no statistically significant warming since 1995. Why bother with the arguments about an El Nino anomaly in 1998? (Incidentally, the red fuzz represents the error ‘bars’.)", when the 1998 El Nino is a major cause of the trend from 1995 being relatively small!!!

    Essentially what Lindzen is doing here is cherry picking and getting in a pre-emptive strike on the argument demonstrating that he has been picking cherries at the same time! [self censored to keep in line with the comments policy ;o)].
    0 0
  46. Journalist at BBC wrote:

    "If a trend meets the 95% threshold, it basically means that the odds of it being down to chance are less than one in 20."


    0 0
  47. Dikran: I share your pain, but not quite as acutely. After all, I actually took (and passed!) a stats class as part of my engineering degree, and I only have a vague understanding of confidence & statistical significance.

    Mind you, some of that might be due to it being close to 20 years since I needed to apply it... :-P

    But if someone who did stats at university level (even if only basic stats) has trouble with statistical significance, what chance those people who struggled with algebra in high school, and have not gone beyond basic financial arithmetic since?
    0 0
  48. Bern, quite. Perhaps Prof. Jones problem is that he does understand it! ;o)
    0 0
  49. Dikran: please do me a favour: come over to 'Carbon Brief' and make the definitive expert comment on this thread. Preferably one that is understandable to the sort of person Bern describes.
    0 0
    Moderator Response: [Dikran Marsupial] Done (I have made a long and detailed post, I'll happily discuss it further if it was over the top ;o)
  50. Denialist "cherrypicking" from this BBC editorial really opened my eyes. Even I know what statistically significant means, but most people don't.

    You did not mention in your article what Pravda and the 9-11 Truther publication "Rock Creek Free Press" said about Dr. Jones.

    They made the same claims as the media you mention. Pravda even cited FOX News as their source.

    Pravda was actually slightly more accurate than the Daily Mail or the Fox News article they cited because they mentioned "statistically significant," but they didn't explain that term.

    The Pravda article was titled "Phil Jones Backs Out Of Global Warming Fuss"
    (2-16-10), but Pravda actually did report:

    The scientist behind the so-called "climate-gate" e-mail scandal now admits there has been no statistically significant global warming since 1995.

    These papers left out important information that Phil Jones told the BBC and that the BBC did report:

    BBC: How confident are you that warming has taken place and that humans are mainly responsible?

    Phil Jones: I'm 100% confident that the climate has warmed. As to the second question, I would go along with IPCC Chapter 9 - there's evidence that most of the warming since the 1950s is due to human activity.
    0 0

Prev  1  2  3  4  Next

You need to be logged in to post a comment. Login via the left margin or if you're new, register here.

The Consensus Project Website


(free to republish)

© Copyright 2019 John Cook
Home | Links | Translations | About Us | Privacy | Contact Us